A real purchasing decision, a real economic tradeoff, and hindsight that conflicts with the original recommendation.
While working at Keyence, I supported a machine builder developing an inspection system for automotive bumper assemblies. The application required verifying color consistency between an ultrasonic sensor face, the bezel surrounding it, and the bumper itself.
Highly reflective surfaces created conditions where lighting, alignment, fixture accuracy, viewing angle, and production variability could significantly influence inspection performance. Silver, white, black, and dark blue combinations were particularly challenging due to reflections and subtle color differences. The project required approximately twenty-four cameras across four production lines.
The machine builder was also evaluating a solution from Cognex. Cognex proposed a more expensive system supported by a formal feasibility study. The Keyence solution was substantially less expensive but had not undergone the same level of validation.
The relationship began as a typical product introduction: demonstrating a newly released Keyence vision system. The price difference immediately attracted attention: the Keyence camera cost approximately $900 CAD, versus roughly $4,000 USD per camera for the Cognex alternative.
The customer provided bezel samples and asked whether the Keyence system could reliably distinguish the color combinations. I tested the samples and demonstrated that it could. The application engineer approved the approach. The mechanical engineer approved the approach. The owner of the machine builder reviewed the solution and approved it after receiving support from the engineering team. The customer was also given a camera to perform additional testing before the final decision.
After evaluation, the machine builder selected Keyence, after reviewing the demonstrations, conducting their own testing, and weighing the economic impact of the alternatives.
Deployment proved more difficult than the original demonstrations. Inspection performance depended heavily on fixture alignment, viewing angle, lighting conditions, and part presentation. The original testing had been performed using supplied samples rather than fully assembled bumper systems operating under production conditions.
I provided multiple training sessions and programming assistance, expecting the customer would become self-sufficient. The customer appeared to have a different expectation. Though never explicitly stated during the sales process, they appeared to expect ongoing programming support from Keyence personnel. Similar expectations emerged with another Keyence representative as portions of the project moved into the United States.
Over time, the customer reported difficulties with the application and eventually paid Cognex to assist with programming and implementation.
The most uncomfortable aspect of this case is that I likely would have made the same recommendation at the time. The demonstrations were successful. The customer had the opportunity to test independently. Multiple stakeholders approved the solution. The economic advantage was substantial. Given the same information and the same experience level, I would likely have viewed the decision as reasonable.
Today I would not. Not because the camera could not perform the task, not because the customer acted irrationally, and not because anyone ignored the available evidence. I would evaluate the application differently because I now place greater weight on production variability, implementation risk, ownership of programming, validation methodology, and long-term maintainability. Today I would likely recommend the Cognex solution despite its higher cost.
What makes this uncomfortable is that the information available at the time was not dramatically different. What changed most was my judgment.
The most interesting aspect of this case is not that the customer selected the lower-cost option. It's how comfortable multiple stakeholders became with the remaining risk. The application engineer approved it. The mechanical engineer approved it. The owner approved it. Additional testing was reportedly performed. A significant cost advantage existed. Yet the production environment contained complexities that were not fully represented during the original demonstrations. The customer did not choose blindly. They chose after review, testing, and multi-stakeholder approval.
Was the decision driven by confidence in the solution, or by confidence that the remaining risk was acceptable relative to the savings?
Customers Buy The Decision They Can Safely Defend™: the purchasing decision was approved by multiple stakeholders, supported by demonstrations, engineering review, customer testing, and substantial cost savings. At the time of purchase, the Keyence solution appears to have been highly defensible despite carrying greater technical and implementation risk. This case provisionally supports the finding, though alternative explanations remain possible.
Confidence Equation™: technical confidence was established through demonstrations and testing. Implementation confidence was largely assumed rather than validated under production conditions. The eventual deployment difficulties suggest these two forms of confidence were evaluated very differently.
Customers Buy The Decision They Can Safely Defend™: this case appears to support the finding at the moment the purchasing decision was made. It does not establish whether the decision remained defensible after implementation difficulties emerged. This suggests a possible boundary condition: the factors that make a decision defensible before purchase may not be the same factors that make it defensible after outcomes become known. This case neither confirms nor refutes that possibility.
Confidence Equation™: implementation confidence was not directly measured before deployment. The conclusion is inferred from subsequent events rather than documented during the original decision process. A stronger test would require evidence of how stakeholders evaluated implementation risk before purchase.
How do organizations decide when cost savings justify accepting additional implementation risk?
This account is based primarily on the salesperson's observations and recollections. The customer's internal decision-making process, risk assessment, and testing activities are not independently documented. Conclusions regarding customer motivations should be treated as informed interpretations rather than verified facts.
The information available at the time was not dramatically different from the information available today. What changed most was judgment. Does experience change how people evaluate risk, or does it simply change which risks they notice?